Annals of Botany 86: 953±962, 2000
doi:10.1006/anbo.2000.1263, available online at http://www.idealibrary.com on
The Chloroplast and Leaf Developmental Mutant, pale cress, Exhibits Light-conditional
Severity and Symptoms Characteristic of its ABA De®ciency
D AV I D R . H O L D I N G * {{, PAT R I C I A S . SP R I N G E R { and S H I R L E Y A . CO O M B E R{
{Division of Life Sciences, King's College London, Franklin-Wilkins Building, 150 Stamford Street,
London SE1 8WA, UK and {Department of Botany and Plant Sciences, University of California,
Riverside, CA 92521, USA
Received: 15 May 2000 Returned for revision: 27 June 2000 Accepted: 11 July 2000
The PALE CRESS gene (PAC) is essential for proper chloroplast and leaf development in Arabidopsis thaliana. The
ability of pac mutants to accumulate signi®cantly more chlorophyll when grown in low light conditions than in high
light conditions suggests that carotenoid de®ciency is at least partly responsible for premature cessation of chloroplast
development. In addition to accumulation of low levels of chlorophyll and carotenoid pigments, pac mutants are
abscisic acid (ABA) de®cient and have characteristics which may be explained by this de®ciency. These include
reduced seed viability and, in enclosed growth conditions, increased leaf growth. Plants transformed with an antisense
PAC construct often bear viviparous embryos which may be symptomatic of a de®ciency in ABA. Since carotenoids
are precursors of ABA, a role for PAC in carotenoid biosynthesis is further supported. The nuclear-encoded,
chloroplast-localized PAC protein has been implicated in the maturation of plastid-encoded mRNAs. Thus, PAC may
a€ect the abundance of one or more chloroplast proteins which function in the synthesis or stability of carotenoids.
Using the PROLIFERA gene as a marker for cell division, it is shown that cell division pro®les in the pac shoot apex
are disrupted. pac leaves are relatively normal in size and shape despite the light intensity-induced variability of leaf
# 2000 Annals of Botany Company
cell defects.
Key words: Abscisic acid, carotenoid, chloroplast development, leaf development, organismal theory, PALE CRESS,
PROLIFERA, vivipary.
I N T RO D U C T I O N
pale cress ( pac) was selected for characterization from a
collection of T-DNA transformed lines of Arabidopsis
thaliana displaying aberrant chloroplast development
(Feldmann, 1991). pac is amongst a small group of such
mutants which also show abnormal leaf development and is
thus a useful tool in the investigation of the coordination
between chloroplast and leaf development. The most
striking aspect of the pac mutant phenotype is its pale
yellow or white cotyledons and leaves resulting from low
level (less than 3 %) accumulation of chlorophylls and
carotenoids (Reiter et al., 1994). Morphological analysis of
pac seedlings grown in moderately high light intensity
revealed that the pac plastids undergo the early stages of
development from proplastids. In young pac leaf plastids
only minimal thylakoid stacking is detectable (Reiter et al.,
1994). This may be the result of photodestruction since a
primary function of carotenoids is to dissipate excess light
energy which otherwise results in highly reactive and
destructive forms of chlorophyll (Anderson and Robertson,
1960, reviewed in Oellmuller, 1989). pac mutants form
normal etioplasts which are capable of the early stages of
post-etiolative development (Reiter et al., 1994). Initial
analysis showed that cellular development in young pac
* For correspondence at: Department of Botany and Plant Sciences,
University of California, Riverside, CA 92521, USA. Fax 909 787
4537, e-mail [email protected]
0305-7364/00/110953+10 $35.00/00
leaves can occur normally, but as leaf expansion proceeds
the characteristic mesophyll cell layers become indistinguishable and premature cessation of mesophyll cell
division results in extensive intercellular spaces (Reiter
et al., 1994). Despite the observed abnormalities in leaf
cellular morphology, overall leaf shape remains normal
(Reiter et al., 1994).
Although analysis of the PAC protein has not yielded
information about its possible function, the predicted
N-terminal region has features suggestive of a chloroplast
transit peptide (Keegstra et al., 1989; Von Heijne et al.,
1989). Characterization of the pac-2 allele (Grevelding et al.,
1996) has demonstrated PAC to be imported into the
chloroplast (Meurer et al., 1998). Within the leaf epidermis,
function of the PAC chloroplast transit peptide is speci®c to
guard cell plastids (Tirlapur et al., 1999). The PAC signal
peptide does not function in plastids of other leaf epidermal
cell types or in plastids of the guard cell progenitors,
suggesting that in the epidermis PAC function is speci®c to
mature guard cells. Furthermore, the function of the PAC
transit peptide is not speci®c to chloroplasts and it is
operational in all types of guard cell plastids, including
etioplasts (Tirlapur et al., 1999).
Initial gene expression analysis revealed that expression
of the nuclear chlorophyll a binding protein (CAB) genes is
una€ected in pac mutants, whereas the expression of a
representative plastid encoded gene, PSBA, is reduced
(Reiter et al., 1994). This analysis was extended by Meurer
# 2000 Annals of Botany Company
954
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
et al. (1998) who showed that the abundance of several
speci®c chloroplast encoded transcripts was reduced.
Furthermore, the processing of some chloroplast transcripts
was found to be aberrant, and larger unprocessed mRNA
precursors were detected. These results suggested that PAC
may be directly involved in plastid mRNA processing and
maturation.
The severity of the pac mutant phenotype is dependent
on light intensity (Grevelding et al., 1996; Meurer et al.,
1998). Using chlorophyll ¯uorescence measurements it was
shown that photosystem II function is disturbed and
photosynthetic electron transport is completely interrupted
in pac mutants grown in moderate to high light intensities
(Meurer et al., 1998). However, plants grown in a low light
intensity showed some greening and displayed variable
chlorophyll ¯uorescence indicating weak photosystem II
activity (Meurer et al., 1998).
In this study we show that the carotenoid de®ciency of
pac is accompanied by ABA de®ciency which can explain
several of the pleiotropic aspects of the phenotype. Our
observations of sense and antisense transgenic plants are
also suggestive of over- and under-production of ABA,
respectively, and collectively support the theory that PAC
plays a role in the carotenoid and ABA pathway. We show
that the light-sensitivity of the pac mutant chloroplast
phenotype also applies to the leaf cell defects and that light
intensity rather than leaf age is the major factor determining phenotypic severity.
M AT E R I A L S A N D M E T H O D S
Plant material
The pale cress mutant was isolated from a collection of TDNA mutagenized lines of the Wassilewskija ecotype of
Arabidopsis thaliana (Feldmann, 1991) and was initially
characterized by Reiter et al. (1994). Seeds were germinated
after 2 d vernalization at 48C, on 0.44 % MS salts
supplemented with 2 % sucrose (unless otherwise indicated
in the text) and either 0.8 % agar or 0.35 % phytagel
(Sigma). Plants were grown under 16 h days, 8 h nights
with white light of 10 mmol m ÿ2 s ÿ1 (low light intensity),
80 mmol m ÿ2 s ÿ1 (moderate light intensity) or 150 mmol
m ÿ2 s ÿ1 (high light intensity). Chlorophyll was extracted
and quanti®ed according to Harborne (1984). ABA was
extracted from 15-d-old wild type and pac seedlings grown
in high light intensity on media containing 2 % sucrose and
quanti®ed according to Bray and Beachy (1985).
Leaf microscopy
Leaves were cut into quarters to aid ®xative penetration
and orientation and the apical quarters were ®xed for 3 h in
4 % paraformaldehyde in 100 mM Pipes ( pH 7) containing
0.05 % Tween 20. Samples were rinsed in 100 mM Pipes
( pH 7) and dehydrated in an ethanol series before being
embedded in methyl butyl methacrylate resin (Electron
Microscopy Sciences) which was polymerized overnight at
608C. Microtome sections 2 mm thick were ¯oated on drops
of water, adhered to glass slides on a 508C hot plate and
stained with 1 % toluidine blue.
For analysis of leaf cellular morphology, ®ve rosette
leaves of 25-d-old plants were sectioned for each light
regime. Average blade length of sampled leaves was 14 mm
for wild-type leaves and 8 mm for pac leaves in both high
and low light regimes.
In situ hybridization
In situ hybridization was performed using a digoxigenin
RNA labelling and detection system (Roche) as described
previously (Lincoln et al., 1994). PRL RNA probes were
synthesized as described in Springer et al. (2000). Sense
PRL RNA probes gave no signal above background (not
shown).
Transgenic plant preparation
PAC cDNA fragments were ampli®ed by PCR from a
PAC genomic clone, pC22 (S. Coomber, unpubl. res.) and
blunt end ligated into the binary vector, pROK2. pROK2
was constructed from pBIN19 (Bevan, 1984) by Dr
Christine Raines (Department of Biological Sciences,
University of Essex, UK) and contains the cauli¯ower
mosaic virus 35S promoter and a nopaline synthase 30
terminator. Plasmid DNA was transformed into Agrobacterium, strain LBA4404, by electroporation (Walkerpeach
and Velton, 1994) and transgenic plants were produced
using root explant cocultivation (Koncz et al., 1990).
Regenerating transgenic plants were selected on
50 mg ml ÿ1 kanamycin at all stages.
R E S U LT S
pac pigment de®ciencies
pac mutants accumulate less than 3 % wild-type levels of
chlorophylls and carotenoids (Reiter et al., 1994) resulting
in pale yellow or white plants (Fig. 1A). Low levels of the
coloured carotenoids, lutein, b-carotene, neoxanthin and
violaxanthin were detectable in pac mutants (Reiter et al.,
1994). To address whether these carotenoid de®ciencies
result in photodestruction of chlorophyll, pac mutants were
grown in low light conditions (10 mmol m ÿ2 s ÿ1). Such
plants were greener than when grown in a relatively high
light intensity (150 mmol m ÿ2 s ÿ1) (compare pac mutants in
Fig. 1A and 1B). However, pac mutants grown in dim light
were not as green as wild type plants grown in dim light
(Fig. 1B). These observations were con®rmed by measurements of total chlorophyll accumulation (Table 1). In
addition, pac mutants grown in low light conditions showed
more growth in the absence of supplied sucrose than in high
light conditions (not shown). This is consistent with the fact
that pac mutants have a higher photosynthetic capacity in
low light conditions than in high light conditions (Meurer
et al., 1998).
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
955
F I G . 1. pac mutant light sensitivity. A, 25-d-old pac mutant seedling grown in high light intensity. Wild type plant shown in background. B, 12-dold wild type (left) and pac mutant (right) seedlings grown in low light. C, pac leaf of a high-light grown plant showing rough surface texture. D,
pac leaf of a low-light grown plant showing smooth surface texture. All panels show plants grown in enclosed, tissue culture conditions.
pac seeds show reduced viability and longevity
We investigated germination of pac mutant seeds relative
to pac heterozygotes and wild type seeds in pooled seeds
isolated from pac heterozygotes. pac mutants accounted for
T A B L E 1. Chlorophyll content of pac mutants and wild type
plants grown in high and low light
Mean chlorophyll content (mg g ÿ1 FW)
Sample
Wild type
pac
High light
Low light
897.5 + 23.1
6.0 + 0.8
185.3 + 13.2
42.3 + 3.9
The values are the mean of ®ve separately measured samples +s.d.
only 13.6 % of viable seeds instead of the expected 25 %
when seeds had been stored at room temperature in dark,
aerobic conditions of approx. 40 % humidity for 6 months.
This suggests either that the pac mutation a€ects embryonic
development or initial seed viability, or alternatively that
the pac mutant seeds have reduced longevity relative to
heterozygous and wild type seeds. To address this and the
question of whether or not pac seeds are more dependent on
a supplied carbon source for germination, we sowed seed
from freshly harvested pac heterozygotes on media containing various concentrations of sucrose. Viability of pac seeds
was reduced with respect to pac/‡ heterozygotes and wild
types and the level of applied carbon source had no
signi®cant e€ect on the percentage of germinating pac
seedlings, suggesting that seed energy shortage is not the
reason for reduced viability of pac mutant seed (Table 2).
T A B L E 2. pac mutant seed viability and longevity
Germination of seeds from pac/
‡ heterozygotes
Number of pac/pac counted
Number of pac/‡ and ‡/‡ counted
% pac/pac
6-month-old seeds
2-week-old seeds
2 % sucrose
0 % sucrose
1 % sucrose
2 % sucrose
3 % sucrose
4 % sucrose
203
1292
13.6
59
232
20.3
51
234
17.9
49
199
19.8
65
309
17.4
62
265
19.0
Average ˆ 18.9 %
956
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
However, the average percentage of germinating pac
mutants in seeds from freshly harvested pac heterozygotes
was 18.9 % (expected 25 %) which, when compared to the
percentage of germinating pac mutants in 6-month-old
seeds (13.6 %), suggests that pac seeds have reduced
longevity in addition to the observed initial reduction in
viability when compared with pac heterozygotes or wild
type. Percentages of pac heterozygotes relative to sibling
wild type seeds suggested that viability and longevity of pac
heterozygotes was not reduced (not shown).
Leaf expansion can be greater in pac leaves than in wild type
leaves
We have shown that although pac mutants are able to
accumulate signi®cantly more chlorophyll in low light than
in high light conditions, they do not accumulate chlorophyll
to the same extent as wild type plants grown in low light.
Consequently one may expect that wild type plants grown
under such conditions would be more photoautotrophic
and thus, show more substantial growth than pac. However,
when wild type and pac mutant plants were grown
adjacently on media containing 1 % sucrose in low light,
pac mutants were usually larger than wild type plants. This
di€erence is the result of greater leaf expansion in pac than
in wild type and is shown quantitatively in Table 3.
pac mutants are de®cient in abscisic acid
Low levels of carotenoids in pac mutants (Reiter et al.,
1994) and the fact that ABA is synthesized from carotenoids, coupled with the pac leaf size data, suggested that pac
mutants might also be ABA de®cient. We measured ABA
concentration in extracts from pac mutant seedlings in
comparison with wild type seedlings and found it was
T A B L E 3. Comparison of leaf size in pac mutants and wild
type plants grown in low light
Mean blade
width
Sample
Wild type
pac
Mean blade
length
Mean total leaf
length
mm
2.8 + 0.8
4.7 + 0.7
3.5 + 1.0
6.5 + 0.8
7.3 + 1.6
14.0 + 1.6
Values for each sample are the mean of measurements from 50
rosette leaves of 6-week-old plants +s.d.
reduced to 12 % of wild type levels [ pac mutants: 18.26
(+6.81) ng g ÿ1 FW; wild type: 147.2 (+49.44) ng g ÿ1 FW].
Antisense and sense over-expression of PAC in transgenic
plants
We investigated PAC function by altering expression
levels of the PAC gene. A 1012 base pair fragment
containing the entire PAC open reading frame was inserted
in antisense or sense orientations downstream of the
cauli¯ower mosaic virus 35S promoter in the binary vector
pROK2. Transgenic plants expressing the antisense or sense
PAC genes and also control plants carrying pROK2 were
regenerated from root tissue (Koncz et al., 1990). The
e€ects conferred by each transgene were compared with the
growth of control plants. As described below and as summarized in Table 4, both transgenes resulted in dramatic
reduction in the recovery of T1 plants and, consequently,
we were unable to propagate stably transformed transgenic
lines with which to demonstrate qualitative alterations in
PAC expression levels. However, our observations of
primary regenerating transgenic plants give important
clues to PAC function and are thus presented here. PCR
analysis on plant genomic DNA con®rmed that complete
and correctly oriented transgenes were present in all plants
analysed (not shown). Northern analyses on selected T0
plants showed that all antisense and sense transgene plants
examined showed expression of the transgene (not shown),
but since the endogenous PAC mRNA is expressed at a very
low level (Reiter et al., 1994) it was not possible to detect
alterations in its abundance.
Antisense over-expression of PAC
Transgenic plants expressing the PAC1012 antisense
transgene had a range of di€erent phenotypes. Plants
regenerated from callus with similar vigour to pROK2
control plants and whilst growing on high cytokinin media
usually displayed healthy green foliage. However, after
plants were transferred to root-inducing media (containing
low cytokinin and high auxin concentrations), leaves often
took on a pale yellow or white appearance resembling those
of pac mutants (Fig. 2A). Under identical conditions,
pROK2 control plants continued to grow with a normal
healthy green appearance (Fig. 2B). A fraction of transferred transgenic shoots for both pROK2 control (12/50)
and PAC1012 antisense (17/47) failed to develop
apical dominance and remained as ill-de®ned mixtures
T A B L E 4. Growth characteristics of PAC1012 antisense and PAC1012 sense transgenic plants
Early growth of transgenic shoots
Fraction of vegetative plants displaying pac mutant-like leaves
Fraction of bolting plants displaying pac mutant-like leaves
Fraction of bolting T0 plants producing T1 seed
Germination and viability of T1 seed
PAC1012 antisense
PAC1012 sense
pROK2 control
Vigorous
14/17
19/30
16/30
6/16: viviparous
6/16: 0 % germination
4/16: 510 % germination
Slow
2/10
0/12
9/12
8/9: 0 % germination
1/9: 100 % germination
Vigorous
0/12
0/38
35/38
35/35: 100 % germination
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
of reverted callus and leaves. The majority of such
vegetative PAC1012 antisense plants displayed some pacmutant like leaves (Fig. 2A). In contrast, vegetative control
plants were always green as shown in the example in
Fig. 2B. Bolting PAC1012 antisense transgenic plants
displayed variable numbers of white or sectored, green
and white leaves as indicated in the example in Fig. 2C, and
whilst pROK2 control plants sometimes had some yellow
or senescent leaves, no such similar white, pac-like leaves
were observed (not shown).
957
Of the 47 transferred PAC1012 antisense shoots, 16
plants produced seed-bearing siliques. In six such plants,
seed germination took place viviparously either directly in
the siliques (Fig. 2D) or on the media below following
premature silique dehiscence (Fig. 2E). This was not
observed in any pROK2 control transgenic plants or in
plants transformed with other PAC cDNA based constructs
(not shown), indicating that the observed vivipary is not
an artifact of the tissue culture process. The primary
transformant shown in Fig. 2D had an otherwise normal
F I G . 2. PAC antisense transgenic plants. A, Typical vegetative PAC1012 antisense shoot with pac mutant-like leaves (arrowheads). B, Typical
vegetative control ( pROK2) shoot displaying healthy green leaves. C, Typical bolting PAC1012 antisense plant with white or sectored, green and
white leaves (indicated with arrowheads). D, Viviparous seed germination from a PAC1012 antisense primary transformant which showed no
white leaves. Arrowheads point to seedlings emanating from a dried silique, arrows point to seedlings loosely attached by roots to mother plant. E,
Viviparous seed germination on media surface (arrows) after premature release of seeds from silique. Seedlings show `pac-like' leaves (indicated by
arrowheads) and arose from a primary transformant exhibiting some `pac-like' leaves.
958
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
appearance during growth and the precociously germinating T1 progeny were also of normal appearance. In
contrast, the T1 progeny shown in Fig. 2E, some of which
have partially white leaves, arose from a primary transformant of the type shown in Fig. 2C which showed some
pac-like symptoms. Since viviparous germination took
place during the early stages of the plant drying out, we
were unable to collect any seed from these plants. Seed from
non-viviparous plants was collected and found to be of very
low viability. Six of the ten plants produced seed with 0 %
viability and the remaining four plants produced seed with
less than 10 % viability. Seed from all control plants
germinated normally.
Sense over-expression of PAC
Early regeneration of transgenic shoots harbouring the
PAC1012 sense construct was obviously compromised
compared to control shoots and antisense transgenic shoots
(Table 4). Shoots regenerating from root explants grew
slowly in comparison to transgenic shoots of control and
other constructs and mostly had a yellow or pale green
appearance. The majority of shoots transferred to high
cytokinin, shoot-inducing media produced very few green
leaves but rather continued to grow slowly as pale green
callus with abundant root growth (not shown). The small
fraction of transgenic shoots which produced leaves and
developed apical dominance appeared phenotypically
normal. However, most of these normal looking transgenic
lines (9/12) produced seeds which did not germinate. A
similar observation was not made in the seed collected from
any control transgenic plants.
Cellular development in pac leaves
pac leaves 4±5 mm in length have severely abnormal
cellular morphology (Reiter et al., 1994). Epidermal cells
are larger than in wild type and their irregular shape and
distribution results in undulating epidermal surfaces.
Furthermore, there is a 50 % reduction in the number of
mesophyll cells in pac relative to wild type, increased air
space and no apparent palisade layer. In light of such
abnormalities, it is interesting that we have always observed
the shape of pac leaves to be normal both at the 4±5 mm
stage and at the fully expanded stage. We analysed the
cellular morphology in fully expanded pac leaves such as
those shown in Fig. 1A and observed cellular abnormalities
which were apparently of equal and greater severity to those
described for 4±5 mm leaves. In medial transverse sections
of 8-mm-long pac leaves, epidermal cells were irregular in
size and shape although not visibly more expanded than
those in the wild type, and epidermal surfaces were highly
convoluted. Mesophyll cells were of variable size and shape
and were apparently more dispersed and a distinct palisade
layer was absent (Fig. 3B).
Leaf blades of pac mutants grown in low light were also
of relatively normal plan shape (Fig. 1D), although petiole
length was often increased relative to high light grown pac
leaves. However, such leaves are visibly smoother in surface
texture than pac leaves of plants grown in high light
(compare Fig. 1C and D). Interestingly, the cellular
structure within these leaves was found to be intermediate
between wild type and high light grown pac leaves
(Fig. 3C). Epidermal and mesophyll cells were of a more
regular shape and size. Mesophyll cells were more closely
F I G . 3. pac leaf cell morphology. A±C show toluidine blue stained medial transverse sections of mature leaves; D±E show PROLIFERA in situ
hybridization of medial longitudinal sections through 8-d-old shoot apices. A, High light-grown wild type leaf. B, High light-grown pac leaf. C, Low
light-grown pac leaf. D, High light-grown wild type shoot apex. E, High light-grown pac shoot apex. Cell layers are labelled in A±C where distinct:
ep, epidermis; pa, palisade mesophyll; sm, spongy mesophyll. Bar (50 mm) in A applies to panels A±C and in D applies to panels D and E.
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
packed than in high light grown pac and though not
continuous, a de®ned palisade layer was evident.
In order to relate pac leaf cell division patterns to those
observed in the wild type shoot apex (Pyke et al., 1991), we
performed in situ hybridization using the PROLIFERA
(PRL) gene (Springer et al., 1995) as a probe. PRL is a
member of the MCM gene family whose proteins function
in the initiation of DNA replication. PRL is speci®cally
expressed in populations of dividing cells within the plant
body (Springer et al., 1995, 2000). In wild type shoot apices,
patchy high level PRL expression is observed throughout
new leaf primordia and young expanding leaves (Fig. 3DÐ
youngest three visible leaves and emerging leaf primordium;
Springer et al., 1995, 2000). This patchiness of expression
probably represents an asynchrony of cell cycle status in
populations of dividing cells. Eventually expression is
reduced in the leaf tip (Fig. 3DÐoldest visible leaf), the
region where cells switch from division to expansion (Pyke
et al., 1991). The cell division characteristics of the pac
shoot apex di€er markedly from that of wild type seedlings.
The tightly packed nature of wild type leaf cells throughout
the apical-basal axis is not observed in expanding pac leaves
(Fig. 3E); instead cells become rounded and loosely packed.
This disorganized nature of the cells in the expanding leaf
and the irregularity of leaf thickness become more severe
with distance from the meristem. PRL expression in the pac
shoot apex is very much reduced compared to wild type. No
high level expression is observed and newly initiated leaf
primordia exhibit only low level expression. No basipetal
gradient of PRL expression is observed in the young leaves.
These results show that under high-light conditions,
abnormal cellular development is observed even in very
young leaves. Furthermore, in shoot apices of older (15 d)
seedlings, cellular abnormalities are observed in the youngest leaves and even in the shoot apical meristem, and PRL
expression in such apices is faint (not shown). Previous
results suggested that cellular development of young pac
leaves occurs normally (Reiter et al., 1994). It is clear from
these studies that the severity of the pac cellular phenotype
is more closely correlated with light intensity than with leaf
age. The pac seedlings used in the initial analyses of leaf
cellular phenotypes were grown in moderately high light
intensities (75±95 mmol m ÿ2 s ÿ1), under which young leaf
cell development is relatively normal (Reiter et al., 1994).
Under the higher light intensity used here (150 mmol
m ÿ2 s ÿ1), the cell defects can be traced back to earlier in
development.
DISCUSSION
Evaluation of the signi®cance of pac pigment de®ciencies
Low light intensities can allow photoautotrophic growth
without photodestruction of chlorophyll in plants with
carotenoid de®ciencies induced either by mutation or
herbicide treatment (Bachmann et al., 1973). Our observations and measurements of pac mutants accumulating more
chlorophyll when grown in low light suggest that the low
level accumulation of photoprotective carotenoids (Reiter
et al., 1994) may result in destruction of chlorophyll under
959
intense light. Carotenoid-de®cient plants usually maintain
similar chlorophyll levels and chloroplast ultrastructures to
wild type plants grown under similar low light intensities
(Bachmann et al., 1973; Frosch et al., 1979). However pac
mutants do not accumulate chlorophyll levels approaching
wild type plants grown under these same low light conditions. This may indicate that the mutation also causes a
direct reduction in chlorophyll biosynthesis by a€ecting the
early part of the carotenoid biosynthetic pathway which is
involved in the production of several other essential plant
compounds including chlorophyll phytol side chains ( for
reviews see Bartley and Scolnik, 1995; McGarvey and
Croteau, 1995).
The nuclear-encoded PAC protein is post-translationally
imported into the chloroplast (Meurer et al., 1998). This
rules out the possibility that PAC is involved in the
transcriptional or translational regulation of the carotenoid
biosynthetic enzymes which are also nuclear-encoded
(Bartley and Scolnik, 1995). The results of Meurer et al.
(1998) indicate that PAC may act as a factor which
functions in the maturation and accumulation of plastid
mRNAs. This general role for PAC is consistent with the
pleiotropic nature of the pac mutant phenotype and allows
the proposal of two ways in which levels of the key
carotenoids could be severely reduced in pac mutants.
Firstly, it is possible that plastid-encoded proteins exist
which function in maintaining the stability or activity of
one or more carotenoid biosynthetic enzymes. Secondly, the
stability of the carotenoids themselves may depend on
plastid-encoded proteins. If the accumulation of such
proteins is reduced, as would be predicted in pac mutants
if PAC has a general defect in plastid mRNA accumulation
and maturation, the result would be a reduced level and not
complete absence of carotenoid pigments.
Characteristics of the pac mutant are consistent with the
observed ABA de®ciency
An intriguing observation was that of the increased leaf
size of pac mutants when grown in enclosed vessels in low
light in the presence of sucrose. If ABA de®cient mutants
are grown in enclosed conditions in which water loss is
minimized, they can be induced to achieve greater size and
faster growth than wild type plants (Trewavas and Jones,
1991). This observation is consistent with the fact that ABA
is a general growth inhibitor and also parallels our
observations of the ABA de®cient pac mutants in which
leaf expansion can be greater than in wild type.
It is well established that ABA is integral to a plant's
adaptation to water de®cit by reducing leaf water conductance, restricting leaf expansion and increasing water
uptake by the roots (reviewed in Hartung and Davies,
1991). ABA de®cient mutants are unable to increase ABA
synthesis in response to dehydration and thus are unable to
close their stomata. They are able to reduce their water
conductance when ABA is applied (e.g. Tal and Imber,
1970). We noted that even in the absence of applied sucrose,
pac mutants survive, albeit with limited growth, for several
weeks when grown in enclosed tissue culture vessels. This is
in sharp contrast to pac mutants grown in soil which dry
960
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
out and die at the two leaf stage and may suggest that
desiccation rather than lack of photosynthesis is more
immediately detrimental under conditions of ¯uctuating
humidity. The reduced capacity of pac mutants to
synthesize ABA suggests that their ability to regulate the
stomatal aperture eciently is likely to be impaired. This
may not be as detrimental to the plant in enclosed humid
conditions as it would be in open conditions of ¯uctuating
humidity.
PAC has recently been implicated to have a direct role in
stomatal guard cell function. Using a translational fusion
between the N-terminal region of PAC and the green
¯uorescent protein, Tirlapur et al. (1999) showed that the
PAC chloroplast transit peptide directs guard cell speci®c
PAC chloroplast translocation. The PAC signal peptide
does not function in plastids of other leaf epidermal cell
types or in plastids of guard cell progenitors, suggesting
that PAC function in the epidermis is speci®c to mature
guard cells. Furthermore, the function of the PAC transit
peptide is not speci®c to chloroplasts and it is operational in
all types of guard cell plastids including etioplasts. This
light- and photosynthesis-independent PAC plastid import
is consistent with a key role for PAC in ABA-mediated
guard cell function.
ABA is reported to be involved in the determination of
leaf identity in some heterophyllous aquatic plants (Golbier
and Feldman, 1989; Ueno, 1998). Such plants display
marked anatomical di€erences between submerged and
aerial leaves. Submerged leaves have no cuticle, they do not
have distinct palisade and spongy mesophyll and they
rapidly dry out if they are abruptly exposed to air. Similar
to terrestrial plants, the involvement of ABA has been
demonstrated during the osmotic stress response of
submerged leaf types (e.g. Young et al., 1987; Golbier
and Feldman, 1989). ABA induces the switch to aerial leaf
type, including the development of properly ordered
palisade and spongy mesophyll. ABA may also be involved
in the ®nal di€erentiation of palisade and spongy mesophyll
in terrestrial plants. Thus, reduced leaf ABA levels in pac
mutants may also be partly responsible for their improper
leaf mesophyll development. Other photosynthetically
de®cient albino mutants exhibit normal leaf cellular
development (e.g. Olive: Hudson et al., 1993; CLA1:
Mandel et al., 1995) suggesting that energy de®ciency
alone may be insucient to cause cellular abnormalities of
the type seen in pac.
The ABA content in developing seeds frequently follows
a common pattern of accumulation (Black, 1991). After
fertilization, ABA levels are very low and rise to a
maximum approximately half-way through seed development. ABA then falls again to very low levels either slightly
before or after initiation of seed desiccation. The peak in
ABA is necessary to prevent germination during seed
development as well as for the acquisition of desiccation
tolerance. The decline in ABA coincides with the maturation of the seed when its presence is no longer required to
maintain dormancy (as reviewed by Black, 1991). pac
mutant seeds showed signi®cantly reduced viability with
respect to wild type and pac/‡ heterozygous seeds. It is
possible that this is the result of reduced ABA accumulation
in the late stages of embryogenesis and the failure to acquire
desiccation tolerance in a proportion of pac seeds. Since the
pathway producing the carotenoids and ultimately ABA is
not completely blocked, it is probable that low levels of
embryonic ABA in addition to maternally provided ABA
from the heterozygous mother plant may be sucient to
allow most developing pac seeds to acquire desiccation
tolerance and to prevent viviparous germination in all pac
mutant seeds. Other aspects of the pac mutation may contribute to the observed reduction in viability. pac embryos
remain white in colour during embryogenesis and can be
easily distinguished from green wild type and heterozygous
sibling embryos. Low levels of carotenoids in pac mutant
embryos may result in oxidative damage to the developing
embryo due to photodestruction of chlorophyll. The white
non-photosynthetic nature of pac embryos may also a€ect
carbohydrate accumulation in embryonic tissues. The role
of carbohydrates in desiccation tolerance and seed longevity is unclear, but their involvement has been implicated
in several ways (Ooms et al., 1993). As well as the possibility
of slightly reduced desiccation tolerance in pac seeds, pac
seeds may not desiccate completely. A feature of other ABA
de®cient or insensitive mutants is that they do not
dehydrate (Black, 1991). If pac embryos are defective in
their desiccation, this may explain the observed reduction in
seed longevity.
Characteristics of PAC transgenic plants are also consistent
with pac ABA de®ciency
The observations of transgenic plants with antisense or
sense overexpression of the PAC gene are also consistent
with its involvement in carotenoid synthesis. They also
support the fact that higher plants synthesize abscisic acid
from carotenoids (reviewed in Taylor, 1991; Zeevaart et al.,
1991). One symptom of ABA de®ciency is that of viviparous seed germination. Some viviparous mutants are
blocked in the early stages of carotenoid biosynthesis
(Moore and Smith, 1985; Neill et al., 1986) and display
white photobleached leaves when grown in normal light
intensities. Other viviparous mutants are defective either
late in the carotenoid pathway or in the committed part of
ABA synthesis and are not visibly de®cient in pigments
(reviewed in Giraudat et al., 1994). The viviparous germination of T1 seeds of PAC1012 antisense transgenic plants
is likely to be the result of reduced seed ABA accumulation
in the latter stages of seed development. Only mildly
a€ected transgenic plants were able to set seed and thus
exhibit vivipary. Severely a€ected antisense lines, in
common with pac mutants, are not able to set seed. By
partially reducing PAC function we have observed a key
e€ect of the pac mutation which is not detectable in pac
mutants.
Several characteristics of PAC1012 sense transgenic
plants are consistent with the suggestion that overexpression or misexpression of PAC can result in increased ¯ux
through the carotenoid pathway which could ultimately
result in raised seed ABA levels and/or loss of spatial and
temporal control of ABA synthesis within the seed. If ABA
levels remained high in the dormant seed, this may result in
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
its inability to break dormancy. This may explain the lack
of germination observed in PAC1012 sense T1 seeds.
ABA is also involved in the control of vegetative aspects
of plant development (Trewavas and Jones, 1991). Root
growth is usually promoted by exogenously applied ABA
(Robertson et al., 1990; Saab et al., 1990) and the root/
shoot ratio is generally increased (Trewavas and Jones,
1991). ABA generally acts as a growth inhibitor in the aerial
part of the plant; leaf initiation, cell division and cell
expansion are all decreased (Trewavas and Jones, 1991).
Regeneration of PAC1012 sense transgenic plants was
considerably delayed compared with transgenic shoots of
control and other constructs. Buds exhibited small, pale
green leaves and abundant root growth. These characteristics are consistent with the suggestion that overexpression
of PAC results in overproduction of ABA.
pac leaves are of normal shape despite defects in leaf cell
division patterns
Cellular abnormalities in pac leaves result in the lack of
de®ned mesophyll layers and undulating epidermal surfaces.
We were intrigued that the overall shape of pac leaves was
always relatively normal despite the variability in severity of
the leaf cell phenotype. We have shown that the severity
depends more on light intensity than on leaf age. The
normal pattern of prolonged cell division in leaf basal
regions after cessation in apical regions, using the criterion
of PRL expression, is not observed in young pac leaves.
Instead, low level PRL expression is detectable in a more
generalized pattern throughout the leaf axis. Thus, the
normal cell division pattern is altered along with cell shape
and number, but overall leaf form remains relatively normal.
These observations in pac mutants are consistent with
theories suggesting that control of leaf growth is e€ected by
yet to be identi®ed mechanisms which act on a whole organ
(organismal) level. Such theories propose that cell division
and expansion are consequential rather than causal in leaf
enlargement (e.g. Haber, 1962; Smith et al., 1996; reviewed
in Kaplan and Hagemann, 1991; Smith, 1996).
AC K N OW L E D E G M E N T S
DRH thanks Elizabeth Bray and Meena Moses for help
with ABA analysis, Kathleen Eckard and Matt Geisler for
technical assistance and Elizabeth Bray for critical comments on the manuscript. This work was funded by the
Biotechnology and Biological Sciences Research Council
(BBSRC) of the UK.
L I T E R AT U R E C I T E D
Anderson IC, Robertson DS. 1960. Role of carotenoids in protecting
chlorophyll from photodestruction. Plant Physiology 35: 531±534.
Bachmann MD, Robertson DS, Bowen CC, Anderson IC. 1973.
Chloroplast ultrastructure in pigment de®cient mutants of Zea
mays under reduced light. Journal of Ultrastructure Research 45:
384±406.
961
Bartley GE, Scolnik PA. 1995. Plant carotenoids: pigments for photoprotection, visual attraction and human health. Plant Cell 7:
1027±1038.
Bevan M. 1984. Binary Agrobacterium vectors for plant transformation. Nucleic Acids Research 12: 8711±8721.
Black M. 1991. Involvement of ABA in the physiology of developing
and mature seeds. In: Davies WJ, Jones HG, eds. Abscisic acid,
physiology and biochemistry. Oxford: Bios Scienti®c Publishers,
99±124.
Bray EA, Beachy RN. 1985. Regulation by ABA of beta-conglycinin
expression in cultured developing soybean cotyledons. Plant
Physiology 79: 746±750.
Feldmann KA. 1991. T-DNA insertional mutagenesis in Arabidopsis:
mutational spectrum. Plant Journal 1: 71±82.
Frosch S, Jabben M, Berg®eld R, Kleinig H, Mohr H. 1979. Inhibition
of carotenoid biosynthesis by the herbicide SAN-9789 and its
consequences for the action of phytochrome on plastidogenesis.
Planta 145: 497±505.
Giraudat J, Parcy F, Bertauche N, Gosti F, Leung J, Morris P, BouvierDurand M, Vartanian N. 1994. Current advances in abscisic acid
action and signaling. Plant Molecular Biology 26: 1557±1577.
Golbier TE, Feldman LJ. 1989. Osmotic stress, endogenous abscisic
acid and the control of leaf morphology in Hippuris vulgaris L.
Plant, Cell and Environment 12: 163±171.
Grevelding C, Suter-Crazzolara C, von Menges A, Kemper E, Masterson
R, Schell J, Reiss B. 1996. Characterization of a new allele of
PALE CRESS and its role in greening in Arabidopsis thaliana.
Molecular and General Genetics 251: 532±541.
Haber AH. 1962. Non-essentiality of concurrent cell divisions for
degree of polarization of leaf growth I. Studies with radiation
induced mitotic inhibition. American Journal of Botany 49:
583±589.
Harborne JB. 1984. Phytochemical methods. London and New York:
Chapman and Hall, 214±218.
Hartung W, Davies WJ. 1991. Drought-induced changes in physiology
and ABA. In: Davies WJ, Jones HG, eds. Abscisic acid, physiology
and biochemistry. Oxford: Bios Scienti®c Publishers, 53±62.
Hudson A, Carpenter R, Doyle S, Coen ES. 1993. Olive: A key gene
required for chlorophyll biosynthesis in Antirrhinum majus.
EMBO Journal 12: 3711±3719.
Kaplan DR, Hagemann W. 1991. The relationship of cell and organ in
higher plants. Bioscience 41: 693±703.
Keegstra K, Olsen KJ, Theg SM. 1989. Chloroplastic precursers and
their transport across the envelop membranes. Annual Review of
Plant Physiology and Plant Molecular Biology 40: 471±501.
Koncz C, Mayerhofer R, Koncz-Kalman Z, Nawrath C, Reiss B, Redei
GP, Schell J. 1990. Isolation of a gene encoding a novel chloroplast protein by T-DNA tagging in Arabidopsis thaliana. EMBO
Journal 9: 1327±1346.
Lincoln C, Long J, Yamaguchi J, Serikawa K, Hake S. 1994. A
knotted1-like homeobox gene in Arabidopsis is expressed in the
vegetative meristem and dramatically alters leaf morphology when
expressed in transgenic plants. Plant Cell 6: 1859±1876.
McGarvey DJ, Croteau R. 1995. Terpenoid metabolism. Plant Cell 7:
1015±1026.
Mandel MA, Feldmann KA, Rocha-Sosa M, Leon P. 1995. CLA1, a
novel gene involved in chloroplast development is highly conserved in evolution. Plant Journal 9: 949±958.
Meurer J, Grevelding C, Westho€ P, Reiss B. 1998. The PAC protein
a€ects maturation of speci®c chloroplast mRNAs an Arabidopsis
thaliana. Molecular and General Genetics 258: 342±351.
Moore R, Smith JD. 1985. Graviresponsiveness and abscisic acid
content of roots of carotenoid-de®cient mutants of Zea mays. L.
Planta 164: 126±128.
Neill SJ, Horgan R, Parry AD. 1986. The carotenoid and abscisic
content of viviparous kernels and seedlings of Zea mays. L. Planta
169: 87±96.
Oellmuller R. 1989. Photooxidative destruction of chloroplasts and its
e€ect on nuclear gene expression and extraplastidic enzyme levels.
Photochemistry and Photobiology 49: 229±239.
Ooms J, Leon-Kloosterziel K, Bartels B, Koorneef M, Karssen C. 1993.
Acquisition of desiccation tolerance and longevity in seeds of
Arabidopsis thaliana. Plant Physiology 102: 1185±1191.
962
Holding et al.ÐChloroplast and Leaf Developmental Mutant, pale cress
Pyke KA, Marrison JL, Leech RM. 1991. Temporal and spatial
development of the cells in the expanding ®rst leaf of Arabidopsis
thaliana (L.) Heynh. Journal of Experimental Botany 42:
1407±1416.
Reiter RS, Coomber SA, Bourett TM, Bartley GE, Scolnik PA. 1994.
Control of leaf and chloroplast development by the Arabidopsis
gene PALE CRESS. Plant Cell 6: 1253±1264.
Robertson JM, Hubick KT, Yeung EC, Reid DM. 1990. Developmental
responses to drought and abscisic acid in sun¯ower roots. I. Apical
growth, apical anatomy, and osmotic adjustment. Journal of
Experimental Botany 41: 325±337.
Saab IN, Sharp RE, Pritchard J, Voetburg GS. 1990. Increased
endogenous abscisic acid maintains primary root growth and
inhibits shoot growth of maize seedlings at low water potential.
Plant Physiology 93: 1329±1336.
Smith LG. 1996. What is the role of cell division in leaf development?.
Seminars in Cell and Developmental Biology 7: 839±848.
Smith LG, Hake S, Sylvester AW. 1996. The tangled1mutation alters
cell division orientations throughout maize leaf development
without altering leaf shape. Development 122: 481±489.
Springer PS, McCombie WR, Sundaresan V, Martienssen RA. 1995.
Gene trap tagging of PROLIFERA, an essential MCM2-3-5-like
gene in Arabidopsis. Science 268: 877±880.
Springer PS, Holding DR, Groover A, Yordan C, Martienssen RA.
2000. The essential Mcm7 protein PROLIFERA is localized to the
nucleus of dividing cells during the G1 phase and is required
maternally for early Arabidopsis development. Development 127:
1815±1822.
Tal M, Imber D. 1970. Abnormal stomatal behaviour and hormonal
imbalance in ¯acca, a wilty mutant of tomato. II Auxin and
abscisic acid-like activity. Plant Physiology 46: 373±376.
Taylor IB. 1991. Genetics of ABA synthesis. In: Davies WJ, Jones HG,
eds. Abscisic acid, physiology and biochemistry. Oxford: Bios
Scienti®c Publishers, 23±38.
Tirlapur UK, Dahse I, Reiss B, Meurer J, Oellmuller R. 1999. Characterization of the activity of a plastid-targeted green ¯uorescent
protein in Arabidopsis. European Journal of Cell Biol. 78: 233±240.
Trewavas AJ, Jones HG. 1991. An assessment of the role of ABA in
plant development. In: Davies WJ, Jones HG, eds. Abscisic acid,
physiology and biochemistry. Oxford: Bios Scienti®c Publishers,
169±188.
Ueno O. 1998. Induction of Kranz anatomy and C4-like biochemical
characteristics in a submerged amphibious plant by abscisic acid.
Plant Cell 10: 571±583.
Von Heijne G, Steppuhn J, Herrmann RG. 1989. Domain structure of
mitochondrial and chloroplast targeting peptides. European
Journal of Biochemistry 80: 525±545.
Walkerpeach CR, Velton J. 1994. Agrobacterium-mediated gene
transfer to plant cells: Cointegrate and binary vector systems.
In: Gelvin SB, Schilperoort RA, eds. Plant molecular biology
manual B1, V2. London: Kluwer Academic Publishers, 1±19.
Young JP, Dengler NG, Horton RF. 1987. Heterophylly in Ranunculus
¯abellaris. E€ect of ABA on leaf anatomy. Annals of Botany 60:
117±125.
Zeevaart JAD, Rock CD, Fantauzzo F, Heath TG, Gage DA. 1991.
Metabolism of ABA and its physiological implications. In: Davies
WJ, Jones HG, eds. Abscisic acid, physiology and biochemistry.
Oxford: Bios Scienti®c Publishers, 39±52.